Experimental Insights into the Thermal Dehydrogenation of Ethylene

Article pubs.acs.org/EF

Experimental Insights into the Thermal Dehydrogenation of Ethylene Diamine Bisborane Using Allyl-Based Ionic Liquids Basudhrity Banerjee, G. Pugazhenthi, and Tamal Banerjee* Department of Chemical Engineering, Indian Institute of Technology, Guwahati, Assam 781039, India ABSTRACT: This work reports the use of allyl-based imidazolium cations for dehydrogenation of ethylene diaminebisborane (EDAB) at three different temperatures, namely, 95, 105, and 115 °C, under vacuum. The allyl-based ionic liquid (IL) was selected by using the infinite dilution activity coefficient (IDAC) as predicted from the COSMO-SAC (COnductor-like Screening MOdel−Segment Activity Coefficient) model. Based on the results of the COSMO-SAC model, the following allylbased ILs were used for experimentation: 1-allyl-3-methylimidazolium dicyanamide ([AMIM][N(CN)2]), 1-allyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([AMIM][Tf2N]), and 1-allyl-3-methylimidazolium bromide ([AMIM][Br]). The highest amount of hydrogen (3.25 equiv) was measured from the EDAB/[AMIM][Br] system at 115 °C. Gas chromatography was conducted to confirm that the gas released was pure hydrogen. To better understand the reaction mechanism of EDAB dehydrogenation, the Reactive Force Field (ReaxFF) method was employed. Further analyses with 1H and 11B NMR were performed on pure IL and IL/EDAB complexes to reassert the role of IL as a catalyst. Thermogravimetric analysis was also conducted on pure EDAB, pure IL, and EDAB/IL complexes to understand the weight loss phenomenon with respect to rising temperature.

1. INTRODUCTION With the limited and ever-decreasing supply of conventional resources, it becomes essential to look for alternative sources of energy. Though fossil fuels are currently the major source of energy, renewable forms of energy are now being explored that may be self-sustaining. One of the highly investigated topics in this regard is that of hydrogen-liberating moieties. Among all the types of energy production, hydrogen has become the latest subject of interest because of its many attractive features, such as negligible amount of pollutant and byproduct. But at the same time, it has its own share of disadvantages, such as its transportation, which becomes difficult due to the need for cryogenic conditions. Eberle and co-workers,1 in their study, spoke about the chemical and physical storage of hydrogen, especially focusing on hydrides and hydrolytic systems. This has led to the study of chemical hydrides, which act as hydrogen storage materials. On thermolysis, these products act as hydrogen-liberating agents. Various compounds are available that can act as hydrogen-liberating materials, such as metal hydrides (MgH2, Mg2NiH4), chemical hydrides (NaAlH4, LiBH4, Mg(BH4)2), carbon-based materials (carbon nanotubes, active carbons, carbon-derived carbon), and metal−organic hydrides (metal−organic frameworks). Among all of the options, the best bet has been placed on chemical hydrides, as they possess very high capacity and liberate hydrogen at moderate temperatures. One class of such compounds, namely, amine boranes, is known to release a substantial amount of hydrogen gas at temperatures below 200 °C and atmospheric pressure. The simplest of these compounds is NH3BH3 (ammonia borane, or AB), which contains 19.6% hydrogen by weight.2 This compound yields about 14% by weight of hydrogen when heated at around 130 °C. It has the highest hydrogen gravimetric capacity among all amine borane adducts. But on the downside, the decomposition of AB also produces borazine (highly toxic) and ammonia.3−5 Furthermore, the © XXXX American Chemical Society

dehydrogenation of AB has certain limitations: its complete decomposition at high temperature of 500 °C and virtually feasible process of regenerating end products such as hydrazine borane. Efforts to overcome these problems involved making chemical modifications to AB molecules, such as using carbon derivatives of AB molecules. Methyl amine borane (MeAB),6 sec-butyl amine borane (SBAB),7 and ethylene diamine bisborane (EDAB)8 are some of the prominent carbonsubstituted amine boranes. Although these derivatives generate lower amounts of hydrogen after heating, there is no formation of volatile compounds such as borazine as side products in dehydrogenation reactions. EDAB was first synthesized by Kelly and Edwards9 by reacting ethylene diamine with diborane under vacuum. The release of 2 equiv of hydrogen at 110 °C was reported, and with prolonged heating, no volatile impurities were found. The same authors then reported an easier approach to producing EDAB by reacting ethylene diamine dihydrochloride and sodium borohydride in tetrahydrofuran using a high-vacuum apparatus.10 Neiner and co-workers8 had reported a hydrogen release of 10% at a temperature less than 200 °C involving a two-step process. At a temperature range of 100−200 °C, the same group reported an intra- and intermolecular dehydrogenation process of EDAB. Leardini and co-workers11 studied the dehydrogenation mechanism of EDAB in a vacuum as well as under inert gas flow and reported a 4 equiv hydrogen release during the first and second desorption stages. Leardini et al.11 further investigated the decomposition of EDAB at high temperatures up to 1000 °C and reported up to four hydrogen desorption events at temperature below 1000 °C. Their study Received: November 1, 2016 Revised: March 29, 2017 Published: March 29, 2017 A

DOI: 10.1021/acs.energyfuels.6b02823 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

occurred was determined to be 264 °C under a nitrogen atmosphere. Hence, we can assume that there is negligible weight loss due to evaporation within 110 °C, which is the working range for dehydrogenation. Despite their potential utility, there has been no systematic study yet on dehydrogenation reactions in allylimidazolium salts. Thus, this work attempts to study the thermal dehydrogenation behavior of these ILs in EDAB. The obtained residual products are characterized using 1H nuclear magnetic resonance (NMR) and 1 B NMR spectroscopy. To determine the stability of EDAB/IL and IL complexes during the thermolysis process, TGA is also performed. Reactive Force Field (ReaxFF) studies have also been conducted to shed light on the proposed reaction mechanism.

was followed up by Carpenter and co-workers, with an intensive study of the infrared matrix of amino borane and its related compounds was studied. It is well known that dehydrogenation reactions in ionic liquids (ILs) give higher equivalents of hydrogen and require a lower induction period.13,14 In one study, Himmelberger15 found that AB in the presence of IL gave zero induction time, higher reaction rates, and faster reaction time. Hence, the unique property of ILs13,14 makes it a first choice as solvent for dehydrogenation studies. In our earlier work,16 we reported a total of 3.96 and 3.52 equiv of hydrogen from the desorption of EDAB/[BMIM][OAc] and EDAB/[EMIM][OAc], respectively, at 105 °C. Bluhm and co-workers17 studied the release of hydrogen using borane complexes in the presence of an IL, 1-butyl-3-methylimidazolium chloride, and compared the results to analogous solid-state reactions. In another recent work, Sahler18 compared the release using IL to that with neat EDAB and observed a higher release of hydrogen equivalents with IL. However, selection of IL is often a cumbersome job, considering the endless combinations it can provide. Further, for a complete selection, experimental regression needs to be avoided. Hence, in our work, we used the COSMO-SAC (COnductor-like Screening MOdel−Segment Activity Coefficient) model,19 which is a variant of COSMO-RS,20 whereby the solubility of EDAB in IL was predicted. In our previous study, we used acetate-based ILs, as they gave the highest solubility for EDAB.16 The next IL in the line of contention was 1-allyl-3-methylimidazolium cation with anions such as bromide, bis(trifluoromethylsulfonyl)imide, and dicyanamide. In this work, dehydrogenation of EDAB in 1-allyl-3-methylimidazolium cation-based ILs at three different temperatures (95, 105, and 115 °C) under vacuum is examined. The purpose of this research was to evaluate the 1-allyl-3methylimidazolium-based cations as possible catalytic materials for thermal dehydrogenation. This was required since the imidazolium-based ILs possess lower thermal stability and higher viscosity. A higher viscosity is detrimental to the processing and solvent recovery operations. Hence, the addition of an allyl group on the imidazolium cation was attempted in this work to effect a change in melting point (Tg). The allyl-substituted imidazolium salts were found to possess moderate viscosity from reported experimental data. For example, the thermogravimetric analysis (TGA) curve of 1allyl-3-methylimidazolium chloride gave an onset temperature of degradation of about 273 °C, which was slightly higher than that with 1-butyl-3-methylimidazolium chloride ([Bmim][Cl]), 254 °C. 1-Allyl-3-methylimidazolium chloride also gave a lower melting point of 17 °C and a considerably lower viscosity of 685 mPa·s at 30 °C compared to [Bmim][Cl], which gave a melting point of 65 °C and a viscosity of 11 000 mPa·s at 30 °C.21 It was also reported that N-allylcarbazole, which has a chemical structure similar to that of the allyl-based IL, does not get homopolymerized even after heating at 228 °C.21 Thus, a lower viscosity gives a strong plasticizing effect so as to introduce amorphous nature within the ILs but at the same time retain its polarity. The relatively lower melting point and viscosity for 1-allyl-3methylimidazolium halide ILs are attributed to the effective suppression of crystallization of the IL by an allyl group on the N-position.22 The IL, 1-allyl-3-methylimidazolium bromide, is known to be stable at a very high range of temperatures (>230 °C). Further, the temperature at which a 10% weight loss

2. COSMO-SAC MODEL Screening of the IL plays a very important role in selecting the IL best suited for dehydrogenation experiments. The COSMOSAC model was employed for the selection of ILs on the basis of their infinite dilution activity coefficient (IDAC) value. The mechanism of screening of ILs has been well defined by our previous work,16 where EDAB was screened in 35 anions with five families of cations. Hence, this is a continuation from our previous work to explore the allyl-based ILs. The COnductor-like Screening MOdel (COSMO) was proposed by Klamt.20 In such a model, the surface properties are calculated from the induced screening charge densities which develop when a molecule is placed in a perfect conductor. This ensures complete polarization. Due to the perfect screening or polarization by the conductor, the charge densities are placed over the interface of the solute and conductor. The entire charge is distributed over a definite number of segments. The charge is then computed by defining an energy of solvation (Esolv) which takes into account the interaction of the segments with the solute nuclei (Z) (first term of eq 1), solute electron density (second term of eq 1), and interaction within the segments (third term of eq 1), as given below:19b 4πεoE solv =

∑ ∑ ZAqαBAα + ∑ qαCα A

1 + 2

α

α

∑ ∑ qαqβDαβ α

(1)

β

Here, qα represents the screening charge of the αth segment. This is then solved by minimizing the equation with respect to the segments’ charges (qα). The equation takes the following form:19b ⎛ dE solv ⎞ ⎜⎜ ⎟⎟ = ⎝ dq α ⎠

∑ ZABAα + Cα + ∑ qβDαβ = 0 A

β

(2)

This process is also known as “self-consistent field” (SCF) or a COSMO calculation. The COSMO files then store the surface screening charge densities, optimized geometries, and energies. These constitute an important descriptor within the COSMO theory, and a histogram distribution of these charges (which usually lies within −0.03 to +0.03 e/Å2) is known as a “sigma profile”. The Segment Activity Coefficient (SAC)19 part of the model is usually the statistical mechanical framework, which simply calculates the energy required to restore the molecule to its original state (restoring free energy). This energy is also B


Article

Energy & Fuels

Table 1. IDAC and Capacity Values of EDAB at T = 25 °C

given in terms of Gibbs free energy of solvation, which in turn can be related to the activity coefficient:19b ln(xi / Sγi / S) = =

solv ΔGisolv / S − ΔGi / i

kT ΔGi*/ solv − ΔGi*/ solv S i kT

⎛ Vi / i ⎞ ⎟ + ln⎜ ⎝ Vi / S ⎠

(3)

IDAC of EDAB

capacity of EDAB (inverse of IDAC)

[AMIM][Br] [AMIM][N(CN)2] [AMIM][Tf2N]

2.64 × 10−10 2.07 × 10−4 3.36 × 101

3.79 × 109 4.83 × 103 2.97 × 10−2

values). The structures of these materials are shown in Figure 1. Dimethyl sulfoxide-d6 (deuteration degree minimum 99.8%, DMSO)

where the subscripts i/S represent the solute (i) in solvent (S), and the asterisk indicates the pseudo-Gibbs energy of solvation, which is merely the non-translational contribution to ΔG and is measurable. The Gibbs free energy of solvation (ΔG*solv) comprises the electrostatic (ES) and van der Waals (vdW) forces. Activity coefficients are then computed by obtaining the difference in the solvation Gibbs energies in the mixture and individual components, respectively. In such a scenario, the densities and dispersion energies between the two phases are assumed to be equal, and hence the final equation takes the following form:19b ⎛ ΔG*ES − ΔG*ES ⎞ i/S i/i ⎟⎟ + ln(γicomb ) ΔGi*/ Ssolv = ⎜⎜ /S kT ⎝ ⎠

solvent

(4)

comb ln(γi/S )

Here, is the same term as used in UNIQUAC or UNIFAC models. The first term of eq 4 (ΔGi/S *ES − ΔGi/i *ES) is the dominating energy term and is referred to as the restoring free energy; i.e., it becomes zero when the molecule is a perfect conductor. Hence, deviation from a perfect conductor is captured with the calculation of ΔGi/S *ES. A resemblance is devised from a classical NPT ensemble in a molecular dynamics (MD) calculation where an escaping molecule (moli) or a segment (segα) relates the chemical potential of the entire system. Thus, we have19b G(N − moli , P , T ) = G(N , P , T ) − μmol , i G(N − segα , P , T ) = G(N , P , T ) − μseg , α

Figure 1. Structures of the materials used: (a) ethylene diamine bisborane (EDAB), (b) 1-allyl-3 methylimidazolium bromide ([AMIM][Br]), (c) 1-allyl-3-methylimidazolium dicyanamide ([AMIM][N(CN)2]), and (d) 1-allyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([AMIM][Tf2N]).

(5)

So, the total energy or the restoring free energy required will be equal to the sum of all segments or histograms, i.e., ΔGi*/ ES S = N

∑α = 1 μα / S in a mixture S. Details of the activity coefficient and methods are described in our earlier work.16 In our work we have computed eq 3 at infinite dilution, i.e., x→0. The lower the values are from unity, the higher the solubility of the EDAB component in an IL. The initial structures of EDAB and ILs (cations and anions separately) were drawn using MOLDEN freeware. Gaussian 09 was then used to carry out the geometry optimization of the EDAB and ILs at the B3LYP/6-311+G(d) level of theory. Frequency optimization was done using the “freq” keyword in Gaussian 09 to detect the presence of any negative or imaginary frequencies. The final structure obtained after the geometry optimization was further used to generate the COSMO files at the same level of theory, using the keyword “SCRF=COSMORS”.

was used as the deuterated solvent for recording NMR spectra. In order to negate the effect of impurities, the ILs were stored under vacuum at 4.5 × 10−2 mbar at 343 K for 24 h prior to dehydrogenation reactions. 3.2. Experimental Setup. A schematic of experimental setup is given in Figure 2. The whole setup can be divided into two distinct parts. The right side of the setup, including the liquid nitrogen condenser (with stopper) and reservoir, is collectively known as the “gas chamber side”. The left side of the setup, comprising the reactor with ports for input and thermowell, is labeled as the “reactor side”. All the dehydrogenation studies were done under vacuum at around 4.5 × 10−2 mbar. The mixture of EDAB along with IL (15 mg of EDAB and 0.5 mL of IL) was kept in the reactor port and heated at different temperatures for various time intervals on the oil bath. The condenser (GC) was then filled with liquid nitrogen. After every time interval of 20 min, the reactor side was opened and the gas was allowed to pass to the other side of the setup. The main reason for using liquid nitrogen as the condensing agent was to condense the undesirable gas so that it accumulates in the collecting flask (MV, right under the liquid nitrogen condenser). Only hydrogen, being the lightest element, travels to the gas chamber via the gas valve. After a certain predetermined time, the gas valve was closed, resulting in separation of the collected gas from the rest of the setup. The amount of hydrogen gas released (equivalents) was then

3. EXPERIMENTAL SECTION 3.1. Materials. Ethylene diamine bis-borane (EDAB) and the ILs, namely, 1-allyl-3-methylimidazolium bromide (≥97% purity, [AMIM][Br]), 1-allyl-3-methylimidazolium dicyanamide (≥98.5% purity, [AMIM][N(CN)2]), and 1-allyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (≥98.5% purity, [AMIM][Tf2N]), were purchased from Sigma-Aldrich (see Table 1 for their IDAC and capacity C


Article

Energy & Fuels

Figure 2. Schematic diagram of the experimental setup. determined by compressing the gas in the gas buret (GB). This was done by raising the mercury in the mercury vessel (MV) by allowing atmospheric air to enter the system through the air valve (AV). The atmospheric air forces the collected gas to rise up, and it is then compressed in the GB. The rise in mercury level was noted for computing the equivalents of hydrogen by using a mercury leveler (ML). In order to negate water as an impurity, an operating temperature close to the boiling point of water was adopted. Further vacuum ensures that no atmospheric air enters the setup, as the vacuum valves (VV) are kept closed throughout the whole experiment. The hydrogen gas release was confirmed by collecting the gas from the GB using a Hamilton 5.0 mL (22/2′/2) gas syringe. Gas chromatography was conducted using a Bruker 450 gas chromatograph with a thermal conductivity detector (GC-TCD technique) to analyze inorganic gases such as argon, nitrogen, hydrogen, and ammonia. The peak location and detection of evolved gas was further checked by GC analysis of pure hydrogen (99.8%) so as to confirm the presence of hydrogen as evidenced by a similar peak location. 3.3. Analysis. Solution 1H NMR spectra were recorded in DMSOd6 at room temperature on a 600 MHz nuclear magnetic resonance spectrometer (Bruker). Spectra were recorded for 10 samples: EDAB, [AMIM][Br], EDAB/[AMIM][Br] (before reaction), EDAB/ [AMIM][Br] (after reaction), [AMIM][Tf2N], EDAB/[AMIM][Tf2N] (before reaction), EDAB/[AMIM][Tf2N] (after reaction), [AMIM][(N(CN)2], EDAB/[AMIM][N(CN)2] (before reaction), and EDAB/[AMIM][N(CN)2] (after reaction). 1B NMR spectra were recorded for EDAB/[AMIM][Br] (before reaction) and EDAB/ [AMIM][Br] (after reaction) using DMSO-d6 as solvent. Thermogravimetric analysis of EDAB, EDAB/[AMIM][Br], EDAB/[AMIM][Tf2N], and EDAB/[AMIM][N(CN)2] was performed on a Mettler Toledo thermogravimetric analyzer (TGA/SDTA model 851). Samples were heated from 30 to 500 °C in a 60 mL/min flow of N2 at a heating rate of 10 °C/min. TGA of pure ILs [AMIM][Br], [AMIM][Tf2N], and [AMIM][N(CN)2] was also done, using a Netzsch thermogravimetric analyzer (TGA/STA 449 F3 Jupiter) from 25 to 500 °C in a 60 mL/min flow of Ar at heating rate of 10 °C/min.

azolium cation with anions bromide, bis(trifluoromethylsulfonyl)imide, and dicyanamide were performed at three different temperatures: 95, 105, and 115 °C (Table 2). Prior to Table 2. Cumulative Amount of Hydrogen Released by EDAB/IL Complexes at Three Different Temperatures: 95, 105, and 115 °C equiv of hydrogen [AMIM][Br] [AMIM][N(CN)2] [AMIM][Tf2N]

95 °C

105 °C

115 °C

2.08 2.74 2.01

2.59 1.36 2.07

3.25 1.32 2.08

the dehydrogenation experiments, all the ILs were kept under vacuum for 24 h at 353 K to minimize the water content and remove the impurities. 1H NMR spectroscopy was performed to validate the purity of the ILs. The 1NMR spectra for pure ILs are given below in Figure 6a,d,g. The peaks corresponding to the hydrogen atoms within the ILs are correlated and validated. The absence of excess peaks suggests negligible impurity. It should be noted that only the IDACs predicted by the COSMO-SAC model for allyl-based ILs are given here. Other values were already reported in our previous work.16 These values are merely taken as qualitative guidelines. As a thermodynamic property, the lower the IDAC values, the higher the solubility of EDAB in ILs and the higher the capacity. This is true since the dehydrogenation proceeds with the stabilization of the intermediate product, namely, the diammoniate of diborane (DADB), which is known to be highly stable in ILs.11 Since this is usually the rate-determining step, we assume that the IL assists in the stabilization of this reactive intermediate species. Hence, the solubility of EDAB in the IL is explored using the IDAC values. For all experiments, the average pressure on both sides of the setup was maintained at 4.5 × 10−2 mbar. A common trend for all the ILs reflects an absence of any induction period. The total dehydrogenation time was found to be the highest for [AMIM][Br] in comparison to the other ILs. For [AMIM][Br],

4. RESULTS AND DISCUSSION 4.1. Dehydrogenation Experiments with 1-Allyl-3methylimidazolium Cation-Based IL. Based on our COSMO-SAC screening results16−22 (see Table 1), dehydrogenation experiments with EDAB and 1-allyl-3-methylimidD


Article

Energy & Fuels complete dehydrogenation of EDAB at 95 °C took 1100 min (Figure 3). However, as the temperature was increased to 105

Figure 5. Equivalents of hydrogen released from EDAB/1-allyl-3methylimidazolium dicyanamide at 95, 105, and 115 °C.

Figure 3. Equivalents of hydrogen released from EDAB/1-allyl-3methylimidazolium bromide at 95, 105, and 115 °C.

and 115 °C, the EDAB/IL mixture was found to be exhausted earlier, at 900 min, thus releasing more hydrogen. It was also found that the reaction rate as well as the reaction time with [AMIM][Br] decreased as the temperature was increased from 95 to 115 °C. For [AMIM][Tf2N] (Figure 4), the run time was 900 min at 95 °C and 720 min at 105 and 115 °C. As the temperature

production of hydrogen increases with temperature, [AMIM][N(CN)2] shows an opposite trend. The highest amount of hydrogen released was 2.74 equiv at 95 °C, with 1.36 and 1.32 equiv at 105 and 115 °C, respectively. For comparison, pure EDAB, which also does not show any induction period,18 was found to release 2.14 equiv of H2 at 120 °C, which is lower than the 2.74 equiv of H2 at 95 °C for [AMIM][N(CN)2]. As already mentioned from the reported literature, it is the stabilization of the reactive intermediate, i.e., diamino diborane (DADB), which controls the dehydrogenation process. This in turn depends on the basicity of the ILs, obtained from the Kamlet−Taft parameter, which quantifies the acidity and basicity of molecular solvents. Jessop et al. further related it to the yield of hydrogen.23 From their measurements they computed the basicity (β) of several ILs. It was found that ILs with Tf2N anions are generally less polar or susceptible to polarization than those with halogen anions. It is also reported that the β value is strongly controlled by the choice of anion. The average basicity values of Tf2N (0.24) and N(CN)2 (0.53) are significantly below those of halogens such as chloride Cl (0.94). This explains the release of higher equivalents of H2 for [AMIM][Br]. A comparison of the ILs with EDAB could not be made, as the dehydrogenation with pure EDAB was conducted in the presence of air, versus nitrogen in our work. The presence of air makes the quantification of hydrogen difficult owing to the fact that other gases may be produced due to oxidation reactions. For all the dehydrogenation processes, it was observed that two ILs, namely [AMIM][Br] and [AMIM][Tf2N], have a catalytic effect, especially at 95 °C. However, the opposite was observed for the remaining IL, namely [AMIM][N(CN)2]. Further, no induction time was observed in each of the three cases, as opposed to a non-zero induction time for the neat EDAB. Also, higher reaction rates and faster reaction times were observed, in agreement with earlier experiments.11−13 A point worth mentioning is that the dehydrogenation mechanism strongly depends on both basicity and dipole moment.24 It is indeed a strong function of basicity, but as mentioned by Sahler et al.,18 there seems to be a distinct tendency that correlates basicity and hydrogen yield, though there certainly is no strict linear correlation. We attempted to locate the values

Figure 4. Equivalents of hydrogen released from EDAB/1-allyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide at 95, 105, and 115 °C.

increased, we observed that the mixture was exhausted faster and released more equivalents of hydrogen. [AMIM][Tf2N] showed a progressive increase in the equivalents of hydrogen released with temperature. However, the lowest cumulative production of hydrogen was 2.01 equiv at 95 °C, 2.07 equiv at 105 °C, and 2.08 equiv at 115 °C; interestly, it did not breach the amount of hydrogen released from pure EDAB, i.e., 2.14 equiv as reported by Sahler et al.18 The third and the final experiment was conducted with [AMIM][N(CN)2] (Figure 5). A run time of 910 min was required for complete dehydrogenation at 95 °C, and 730 min at 105 and 115 °C. Unlike the other ILs, whose cumulative E


Article

Energy & Fuels Table 3. Rate Constants for Decomposition of EDAB in [AMIM][Br] Using the Avrami−Erofeyev Model

a

T (°C)

β1

β2

n1

n2

k10 × 10−17

k20 × 10−17

E01 (kJ mol−1)

E01 (kJ mol−1)

total H2 (equiv)a

95 105 115

1.11 1.93 1.16

0.97 0.65 2.12

1.02 1.51 1.59

2.24 2.92 2.39

0.68 1.50 7.35

3.90 3.12 7.50

146.99 156.01 162.8

157.48 142.09 136.42

2.08 (2.08) 2.58 (2.59) 3.25 (3.25)

Numbers in parentheses represent the equivalents of hydrogen obtained experimentally.

solve the ordinary differential equations to find the decomposition rate per second. These data were then compared with our experimental data to find the error. Only EDAB/[AMIM][Br] was used for the study, as it gave the highest amount of hydrogen upon thermal decomposition. The error was improved by using the built-in GA subroutine to predict new sets of parameters or variables. The following seven variables were computed using the function: reaction orders for the two reactions (n1 and n2), activation energies (E01 and E02), temperature-independent rate constants (k01 and k02), and equivalents of hydrogen released (β1 and β2). The results are compiled in Table 3. It can be seen that the activation energy for [AMIM][Br] in eq 6 increases, while the opposite is true for eq 7. Hence, an operation at 105 °C seems to be an optimized choice. The total equivalents calculated also agree with the experimental measurement (in parentheses). Thus, it is indeed a kinetic effect, as the IL is seen to pronouncedly lower the activation energies for the rate-determining step, S2 → S3 + β2H2, where S2 is the intermediate. 4.3. 1H NMR Spectra of EDAB with 1-Allyl-3-methylimidazolium Cation-Based IL. 1H NMR spectra were recorded to study and characterize the pure EDAB, pure IL, and EDAB/IL mixture before and after reactions. A comparative discussion of 1H NMR plots of pure EDAB, pure IL, and EDAB/IL complexes will determine the possible formation of any product after dehydrogenation and further elucidate the role of IL in dehydrogenation. Here peaks from the literature27 were used to identify the peak of our compounds. Figure 6a shows the 1H NMR study of pure [AMIM][Br], where the chemical shifts of the hydrogens in the imidazolium ring are identified at 9.35, 7.82, and 7.80 ppm, respectively. The hydrogen atom of the N−CH2 bond is located at 5.29 ppm and that of N−CH3 at 3.87 ppm. The chemical shift at 6.01 ppm is attributed to the C−H bond, while peaks at 4.90 and 4.89 ppm represent CCH2. Figure 6b shows the NMR spectra of EDAB/[AMIM][Br] (before reaction). From the previous 1H NMR study of pure EDAB, we can identify the BH3 peak at 1.19 ppm, CH2 at 2.59 ppm, and NH2 at 5.29 ppm. The remaining chemical shifts of the respective hydrogen atom(s) of the IL remain the same. Figure 6c represents the NMR spectra of EDAB/[AMIM][Br] (after reaction). Here we observe that there is no change in the initial structure of the IL after the dehydrogenation is completed. After the dehydrogenation study of EDAB/[AMIM][Br], it can be concluded that the liberation of hydrogen took place only from EDAB on the basis of the disappearance of peaks of BH3, as evident from Figure 6c. From this study, we can conclude that the IL merely acts as a catalyst since its initial structure is retained even after reaction, whereas the absence of BH3 peaks in EDAB indicates dehydrogenation. In a similar manner, Figure 6d represents the 1H NMR study of pure [AMIM][Tf2N]. The structure of the cation, namely, 1allyl-3-methylimidazolium, remains the same; hence, it is the anion which causes the change in the spectra. As the anion does

for allyl-based ILs from the only known solvatometric properties reported but were unsuccessful in explaining the behavior.23 The only pattern worth mentioning is that the basicity of Tf2N, N(CN)2, and Br follows a decreasing trend, and so does the amount of hydrogen equivalents released, but this is true only for [AMIM][Br]. In this circumstance, the unavailability of the dipole moment or the degree of polarizibility for allyl-based ILs is a hindrance to explaining the unusual trend for [AMIM][Tf2N], which in fact gave a zero activation energy while going from from 95 to 115 °C. Further, besides the mass-transfer effects, thermal effects also had a pronounced role in the dehydrogenation. In a recent work it was also found that [Tf2N]-based ILs do not possess good thermal characteristics.24 Hence, a comprehensive study needs to be performed to determine the exact mechanism. 4.2. Rate Constants for Decomposition of EDAB in [AMIM][Br] Using the Avrami−Erofeyev Model. The hydrogenation was found to be due to the kinetic effect, which also agrees with previous work.11 In order to explain the effects of dehydrogenation, we regressed the experimental kinetic data using the Avrami−Erofeyev model25 to obtain the activation energy, rate constant, and reaction order. In our earlier work26 we demonstrated that the decomposition of AB in the presence of IL is a two-step reaction. The Avrami− Erofeyev model is used since it is known to predict reaction orders of more than unity.20 For the EDAB + IL mixture, the reaction can be expressed as

S1 → S2 + β1H 2

(6)

S2 → S3 + β2 H 2

(7)

dα1 = n1k1(1 − α1)[−ln(1 − α1)]n1− 1/ n1 dt

(8)

dα 2 = n2k 2(1 − α2)[−ln(1 − α2)]n2 − 1/ n2 dt

(9)

k1 = k10 e E01/ RT

(10)

k 2 = k 20 e E02 / RT

(11)

Here αi is the extent of decomposition for eqs 6 and 7, respectively, while S1 stands for EDAB, S2 for the intermediate DADB, and S3 for the oligomeric or the residual products. These correspond to β1 and β2 equivalents of hydrogen, respectively. ni and ki represent the order and rate constant for the ith reaction, whereas E01 and E02 represent the activation energies for reactions 6 and 7, respectively. In order to regress the equations, we used the gamultiobj function in MATLAB. Using this function, a set of unknowns can be found iteratively by setting bounds for each parameter. So in our case, at each iteration we computed two quantities, (a) the cumulative hydrogen production from EDAB + IL mixtures and (b) the corresponding decomposition rate per second. The Adams−Bashford−Moulton method was used to F


Article

Energy & Fuels

Figure 6. continued

G


Article

Energy & Fuels

Figure 6. 1H NMR spectra: (a) pure [AMIM][Br], (b) EDAB/[AMIM][Br] (before reaction), (c) EDAB/[AMIM][Br] (after reaction), (d) pure [AMIM][Tf2N], (e) EDAB/ [AMIM][Tf2N] (before reaction), (f) EDAB/ [AMIM][Tf2N] (after reaction), (g) pure [AMIM][N(CN)2], (h) EDAB/ [AMIM][N(CN)2] (before reaction), and (i) EDAB/[AMIM][N(CN)2] (after reaction).

with a reduced peak area after dehydrogenation, indicating a partial release of hydrogen. The mechanism of [AMIM][N(CN)2] here needs further investigation, as it might act as a regenerating catalyst because of the basicity of the dicyanamide anion.28 The solubility trait of the dicyanamide anion depends on its donor ligand property.29 4.4. 11B NMR Spectra of EDAB with 1-Allyl-3-methylimidazolium Bromide. 1H NMR only gives insights into the shielding and deshielding effects of the hydrogen atoms in EDAB/IL complexes. Hence, a need is also felt to understand the boron moieties within EDAB. 11B NMR spectra were recorded on samples both before the reaction and after the reaction. This has been performed using the IL [AMIM][Br] due to its having the highest hydrogen release. Further, it should be noted that the 1-allyl-3-methylimidazolium cationbased ILs do not possess boron species, which in turn will not interfere with the boron atoms in EDAB. Figure 7a represents the 11B NMR spectrum of EDAB/ [AMIM][Br] before reaction. The sp3 −BH3 functional group can be assigned to the chemical shift at −19.83 ppm, which agrees with the literature.8 After dehydrogenation, a chemical

not possess hydrogen atoms, the spectra will be similar to those of [AMIM][Br]. Figure 6e represents the 1H NMR study of EDAB/[AMIM][Tf2N] (before reaction), where the hydrogen peaks of EDAB are shown along with the hydrogen peaks of IL. Figure 6f represents the 1H NMR study of EDAB/[AMIM][Tf2N] (after reaction), where we can see a similar trend; i.e., after dehydrogenation, the BH3 bond in EDAB is absent. Hence, we conclude that EDAB takes part in the dehydrogenation process. Likewise, [AMIM][N(CN)2] shows similar spectra when compared to [AMIM][Br] and [AMIM][Tf2N]. In a similar manner, Figure 6g−i shows 1H NMR spectra of pure [AMIM][N(CN)2], EDAB/[AMIM][N(CN)2] (before reaction), and EDAB/[AMIM][N(CN)2] (after reaction). The 1 H NMR spectra for the residual product gave a complete disappearance of the −BH3 peak for both [AMIM][Br] (Figure 6c) and [AMIM][Tf2N] (Figure 6f). However, the same cannot be said of [AMIM][N(CN)2] (Figure 6i), where −BH3 was visible, while the −CH2 peak was not present. So, in such a case, EDAB may tend to form a cyclic moiety based on an sp2hybridized boron atom which cannot be isolated in 1H NMR, as in Figure 6i. The −NH2 peak is present in all the ILs but H


Article

Energy & Fuels

Figure 7. 11B NMR spectra: (a) EDAB/[AMIM][Br] (before reaction) and (b) EDAB/[AMIM][Br] (after reaction).

shift helps in the formation of an sp2 trigonal boron entity, which is assigned at 20.14 ppm. It is to be noted that the negative side of the spectrum depicts the sp3 BH and BH3 attachment. For dehydrogenation experiments, a peak at −39 ppm is assigned to BH4, which is an intermediate, and at −10

ppm for asymmetric BH2 resonance. However, in the residual product (Figure 7b), a chemical shift of 20.14 ppm is seen. This can be assigned to the sp2 trigonal boron. Similarly, the peak at −1.47 ppm can be assigned to N-BN-N borons resulting from chain-branched polymeric structures.17 Hence, the peaks after I


Article

Energy & Fuels dehydrogenation nullify the possibility of having BH4 and asymmetric BH2 in the final product. This is due to the fact that the intermediate formation is a fast reaction which happens during the course of the overall reaction. This can be detected with in situ measurements, which is beyond the scope of this work. However, for a detailed mechanism, we shall attempt to join the observations with ReaxFF simulations in section 4.6. 4.5. TGA Analysis of EDAB with 1-Allyl-3-methylimidazolium Cation-Based IL. To confirm the thermal stability of the reacting mixture, TGA analysis was carried out on both pure EDAB and EDAB/IL complexes. The experiments were conducted for a wide range of temperatures (25−500 °C), as [AMIM][Tf2N], being a highly stable IL, goes through a weight loss until 455 °C, in comparison to the other two ILs. For [AMIM][Br], the range remains almost similar, but when compared with EDAB/[AMIM][Br] and pure [AMIM][Br], it is seen that the hydrogen is being released earlier for EDAB/ [AMIM][Br]. EDAB/[AMIM][Br] was more thermally stable as a mixture. However, the comparison is only realistic within a temperature range of 25−120 °C, since it is these regions where the hydrogen gets evolved from EDAB. It is also a known fact that the melting point of EDAB varies from 116 to 120 °C. Moreoever, it is difficult to predict whether EDAB catalyzes the decomposition of the allyl-based ILs or vice versa at higher temperatures. Figure 8a represents the TGA profiles of EDAB, EDAB/ [AMIM][Tf2N], EDAB/[AMIM][Br], and EDAB/ [AMIM][N(CN)2].The black line represents EDAB, while the orange, blue, and pink dotted lines represent the TGA profiles of EDAB/[AMIM][Tf2N], EDAB/[AMIM][Br], and EDAB/ [AMIM][N(CN)2], respectively. On considering the weight loss behavior of pure EDAB, we can state that EDAB shows a negligible weight loss until 106 °C, which confirms the melting point of EDAB. EDAB is seen to depict an induction period, as the weight loss remains constant. However, an intensive mass loss is recorded from 106 to 150 °C, where all the EDAB is consumed. On comparison with EDAB/IL complexes, we can see that the major weight loss starts from 250 °C, and goes downward (except for the EDAB/ [AMIM][N(CN)2]), which proves that, in the experimental studies, it is the EDAB that is going through the thermal dehydrogenation process. From the TGA profiles of EDAB and EDAB/IL complexes, the H2 liberation with ILs is immediately observed from the slope of TGA lines with no induction period. This is suppressed by the presence of ILs. If we now analyze the stability of the EDAB/IL complexes, then we can state that the most stable IL is [AMIM][Tf2N], followed by EDAB/[AMIM][Br] and [AMIM][N(CN)2]. When considering EDAB/[AMIM][Tf2N] (orange dotted line), we see that the first mass loss of 12.36% was observed at 368 °C, and the second major mass loss of 80.52% continued from 368 to 465 °C. The residual mass was calculated to be 3.52%. For the EDAB/[AMIM][Br] system (blue dotted line), the first weight loss of 69.87% occurs at 235 °C, and the second major weight loss of 16.96% takes place at 400 °C. After the second major loss, there is a gradual loss of around 1.96% when the sample is heated from 400 to 500 °C. For EDAB/ [AMIM][Br], further minor desorption takes place at 135 °C, with 4.97% mass loss, and the second one at 225 °C with corresponding weight loss of 5.12%. The residual mass was calculated to be 1.12%. On observing the slope of EDAB/ [AMIM][N(CN)2] (pink dotted line), we can see that the stability drops at a higher rate than for the other ILs. Three

Figure 8. (a) TGA profiles of pure EDAB, EDAB/[AMIM][Br], EDAB/[AMIM][Tf2N], and EDAB/[AMIM][N(CN)2]. (b) TGA profiles of pure ILs: [AMIM][Br], [AMIM][Tf2N], and [AMIM][N(CN)2].

consecutive mass losses of 5.15%, 8.76%, and 1.21% were observed at relatively low temperatures of 100, 125, and 150 °C, respectively. The minor losses confirm that the IL ([AMIM][N(CN)2]) is unstable in comparison to the other two ILs. Further three major mass losses for the same IL were observed from 265 to 500 °C; the most significant was found to be 23.59% at 375 °C. TGA studies on EDAB and EDAB/IL mixtures were done to have an understanding about the thermal stability of the reacting mixtures; however, TGA studies of pure ILs were equally important, as they would throw light on the thermal performance. TGA profiling of the pure ILs was done so as to see whether or at what temperature the IL decomposes and how it affects the dehydrogenation process of EDAB. The TGA profiles of EDAB, EDAB/IL mixture, and pure IL are provided separately, as two different instruments were used to record them. Figure 8b represents the TGA profile of IL [AMIM][Br] (black dashed line), where a major mass loss of 83.67% was observed between the temperature range of 260−375 °C, with a residual mass of 9.34%. In the case of [AMIM][Tf2N] (blue dashed line), we observe a similar trend of thermal decomposition, where the one and only major mass loss of 85.23% takes place at a temperature of 350−455 °C. However, like the EDAB/IL mixture, we can also observe the variant J


Article

Energy & Fuels

Figure 9. Proposed reaction mechanism of EDAB dehydrogenation using ReaxFF.

EDAB. For the intermolecular dehydrogenation, the formation of EDAB dimer is the initiation process. As the dehydrogenation proceeds, a number of large chains of polymerized EDAB dimer are formed. The abstraction of hydrogen then takes place from each repeat unit during the formation of the BH2 cyclic group and BN bonds. We have adopted the ReaxFF simulation to investigate the reaction mechanism. ReaxFF30−32 was developed for bond dissociation and formation using MD simulation. The force field parameters are derived from the quantum mechanics (QM) and are directly used for the system under investigation. It is based on the semiempirical interaction potential, where the potential energy of the system is described by different energies of the system. A detailed description of the ReaxFF force field is given in our previous work30 and detailed in an excellent review by van Duin.32 ReaxFF simulation were carried out in the SCM suite.33 Thirty molecules of optimized EDAB molecules were placed randomly inside a periodical box size of 25 × 25 × 25 Å. The force field defined for such a reactive simulation was C/H/O/

behavior of [AMIM][N(CN)2] (red dashed line) where the thermal stability of the IL drops frequently and quickly in comparison to those of the other two ILs. The first major mass loss of 53.69% takes place at a temperature of 255−280 °C, followed by the last desorption of 39.90% at 425 °C, with a residual amount of 6.41% remaining. TGA study furthers confirms that no hydrogen was released from the ILs, as none of the above-described ILs decomposed before 250 °C, and the highest operating temperature for thermal dehydrogenation of EDAB with 1-allyl-3-methylimidazolium cation-based IL was only 115 °C. 4.6. Proposed Reaction Mechanism Using ReaxFF. The reaction mechanism of ethylene diamine bisborane was proposed earlier by Leardini et al.11 and Neiner et al.8 They proposed two possible pathways based on either an intramolecular or intermolecular approach. In the intramolecular dehydrogenation procedure, an absence of oligomeric EDAB structure is observed after the complete dehydrogenation, which further leads to the formation of cyclic dehydrogenated K


Article

Energy & Fuels N/S/B.31,34 The system was first minimized at 10 K at a time step of 0.25 fs. The process is a non-reactive process in which the system is validated with the intermolecular potential so as to reduce any overlap of assembly of atoms. Initially, a number of NVE simulations with different time steps were performed to find the correct time step. The time step which helped in the conservation of energy was adopted for the subsequent minimization and production stages. The equilibration (nonreactive) was then carried out in an NVT ensemble for 12.5 ps with a time step of 0.1 fs. The production run then proceeds with NVT-MD simulation with a time step of 0.25 fs. The temperature of the system was kept at 1000 K, as a hightemperature simulation tends to give computationally affordable results with greater accuracy. The temperature was controlled via a Berendsen thermostat35 with a damping constant of 0.5 ps. The total time was kept at 250 ps to study the dehydrogenation process. After the simulation, the trajectories of all the molecules and intermediates were visualized from the start to the end of EDAB dehydrogenation. The hydrogen was found to evolve from the −BH3 functional end of EDAB (Figure 9). The first abstraction in the hydrogen atom is observed at around 18.18 ps when EDAB releases its hydrogen atom(s). The hydrogen then rejoins it at the other end of the −BH3 group in EDAB. This step is crucial and supports the idea of chain-branched polymerization, which is also indicated by 11B NMR (−1.47 ppm). This step being short-lived, the product quickly takes the form of a polymeric cyclic compound, which may be attributed to the sp2 trigonal boron as per 11B NMR spectra. Thereafter, the hydrogen finally detaches itself from the EDAB entity. The detachment of hydrogen from the rest of the EDAB cluster hence becomes the rate-determining step. This mechanism partially supports the reaction mechanism reported by Leardini et al.11 The reaction then goes on until 250 ps to give the end products formed during the latter part of the experiment. The later end products are not significant because the rate of hydrogen release is highest at the start of the reaction. As observed from the reaction mechanism obtained from the ReaxFF study, it can be confirmed that the release of hydrogen follows a two-step process, which has been mentioned in section 4.2. For the decomposition of EDAB in [AMIM][Br], the following stoichiometry is proposed:

2 (linear molecule of Figure 9). Hence, the equivalents arising at higher temperature are primarily due to eq 1, as it is the cumulative equivalents which are reported in Figures 3−5. The same cannot be said for other ILs, since equivalents of hydrogens keep on increasing with temperature, which implies that the rate of the reaction is endothermic in nature for both eqs 1 and 2. This is reflected when we compare Figure 3 for [AMIM][Br] and Figure 4 for [AMIM][Tf2N].

5. CONCLUSIONS The thermal hydrogenation of EDAB was carried out using ILs as solvent at three different temperatures, 95, 105, and 115 °C. The COSMO-SAC model was used to select the particular ILs for experimental studies from the vast database of cations and anions. Hence, the ILs with the highest IDAC values, namely, 1-allyl-3-methylimidazolium bromide ([AMIM][Br]), 1-allyl-3methylimidazolium dicyanamide ([AMIM][N(CN)2]), and 1allyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([AMIM][Tf2N]), were chosen for experimental studies. It was observed that the presence of IL had a catalytic effect on the thermal dehydrogenation of EDAB. No induction time was observed, as opposed to a non-zero induction time for pure EDAB. The reaction rate and reaction time was also found to increase with temperature. At T = 105 °C, the hydrogen release from the IL-EDAB mixtures was found to decrease in the following trend: EDAB/[AMIM][Br] (2.59) > EDAB/ [AMIM][Tf2N] (2.07) > EDAB/[AMIM][N(CN)2] (1.36). 1 H NMR studies for the residual compounds indicated the catalytic role of IL, which is evident in the reduction of the −NH2 peak. 1B NMR also helps in confirming the presence of sp2 trigonal boron in EDAB after the dehydrogenation reaction. ReaxFF studies have further validated the reaction mechanism pathway for dehydrogenation of EDAB. TGA studies of EDAB, EDAB/IL, and pure IL have also confirmed the thermal stability of each group separately, indicating no mass loss from the pure IL.

■

Corresponding Author

*E-mail: [email protected]. Phone: +91-361-2582266. Fax: +91-361-2582291. ORCID

Δ

C2H14B2N2 → C2H14B2N2 + β1H1

Tamal Banerjee: 0000-0001-8624-6586

(12)

Notes

The authors declare no competing financial interest.

Δ

C2H14B2N2 → C2H12B2N2 + β2 H 2

AUTHOR INFORMATION

■

(13)

ACKNOWLEDGMENTS The work reported in this article was financially supported by a research grant (SB/S3/CE/063/2013) from the Science and Engineering Research Board (SERB), Department of Science and Technology (DST), Government of India. We would also like to thank the Indian Association for the Cultivation of Science and Central instrument facility, IIT Guwahati, for letting us use their facilities.

where C 2 H 14 B 2 N 2 or EDAB represents S 1 , C 2 H 14 B 2 N 2 represents S2, and β1 stands for equivalents of hydrogen released, as given in eq 6. Equation 13 is similar to eq 7, where S3 represents the oligomeric or the residual products. S2 and S3 have the same empirical formulas but represent the successive polymeric cyclic intermediate, as shown in Figure 9. It is interesting to note that the equivalents of hydrogen released for [AMIM][N(CN)2] seems to overlap close to 80 min (Figure 5). It is after this zone that the temperature plays a role. One way to explain it is that, while an endothermic reaction is seen to govern eq 11, an exothermic reaction dominates eq 13. As solubility is known to be a sum of entropic and energetic forces, such a behavior can be captured by recording the excess enthalpy data for [AMIM][N(CN)2]− EDAB mixtures, which is not measured here. So this phenomenon corresponds to the enhanced formation of step

■

REFERENCES

(1) Eberle, U.; Felderhoff, M.; Schüth, F. Chemical and Physical Solutions for Hydrogen Storage. Angew. Chem., Int. Ed. 2009, 48, 6608−6630. (2) Peng, B.; Chen, J. Ammonia borane as an efficient and lightweight hydrogen storage medium. Energy Environ. Sci. 2008, 1, 479−483.

L


Article

Energy & Fuels

(23) Jessop, P. G.; Jessop, D. A.; Fu, D.; Phan, L. Solvatochromic parameters for solvents of interest in green chemistry. Green Chem. 2012, 14, 1245−1259. (24) Wadekar, V. V. Ionic liquids as heat transfer fluids−An assessment using industrial exchanger geometries. Appl. Therm. Eng. 2017, 111, 1581−1587. (25) Ahluwalia, R. K.; Peng, J. K.; Hua, T. Q. Hydrogen release from ammonia borane dissolved in an ionic liquid. Int. J. Hydrogen Energy 2011, 36, 15689−15697. (26) Mahato, S.; Banerjee, B.; Pugazhenthi, G.; Banerjee, T. Optimization and quantum chemical predictions for the dehydrogenation kinetics of Ammonia Borane-Ionic Liquid mixtures. Int. J. Hydrogen Energy 2015, 40, 10390−10400. (27) Furukawa, S.; Inoue, N.; Ishioka, T.; Furuya, K.; Harata, H. Rapid Decomposition of Cellulose Dissolved in Ionic Liquid Using Gas−Liquid Interface Discharge. Jpn. J. Appl. Phys. 2012, 51, 070205. (28) Forsyth, S. A.; MacFarlane, D. R.; Thomson, R. J.; von Itzstein, M. Rapid, clean, and mild O-acetylation of alcohols and carbohydrates in an ionic liquid. Chem. Commun. 2002, 7, 714−715. (29) MacFarlane, D. R.; Golding, J.; Forsyth, S.; Forsyth, M.; Deacon, G. B. Low viscosity ionic liquids based on organic salts of the dicyanamide anion. Chem. Commun. 2001, 16, 1430−1431. (30) Bhoi, S.; Banerjee, T.; Mohanty, K. Molecular dynamic simulation of spontaneous combustion and pyrolysis of brown coal using ReaxFF. Fuel 2014, 136, 326−333. (31) Weismiller, M. R.; Van Duin, A. C. T.; Lee, J.; Yetter, R. A. Reactive Force Field Development and Applications for Molecular Dynamics Simulations of Ammonia Borane Dehydrogenation and Combustion. J. Phys. Chem. A 2010, 114, 5485−5492. (32) Van Duin, A. C. T.; Dasgupta, S.; Lorant, F.; Goddard, W. A., III ReaxFF: A Reactive Force Field for Hydrocarbons. J. Phys. Chem. A 2001, 105, 9396−9409. (33) Scientific Computing & Modelling (SCM). ADF; SCM, Theoretical Chemistry, Vrije Universiteit: Amsterdam, The Netherlands, 2013; http://www.scm.com. (34) Kamat, A. M.; Van Duin, A. C. T.; Yakovlev, A. Simulations of Laser-Induced Incandescence of Soot Using an Extended ReaxFF Reactive Force Field. J. Phys. Chem. A 2010, 114, 12561−12572. (35) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola, A.; Haak, J. R. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 1984, 81, 3684−3690.

(3) Stephens, F. H.; Pons, V.; Tom Baker, R. Ammonia−borane: the hydrogen source par excellence? Dalton Trans. 2007, 2613−2626. (4) Al-Kukhun, A.; Hwang, H. T.; Varma, A. A Comparison of Ammonia Borane Dehydrogenation Methods for Proton-ExchangeMembrane Fuel Cell Vehicles: Hydrogen Yield and Ammonia Formation and Its Removal. Ind. Eng. Chem. Res. 2011, 50, 8824− 8835. (5) Halseid, R.; Vie, P. J. S.; Tunold, R. Effect of ammonia on the performance of polymer electrolyte membrane fuel cells. J. Power Sources 2006, 154, 343−350. (6) Jaska, C. A.; Temple, K.; Lough, A. J.; Manners, I. Transition Metal-Catalyzed Formation of Boron-Nitrogen Bonds: Catalytic Dehydrocoupling of Amine-Borane Adducts to Form Aminoboranes and Borazines. J. Am. Chem. Soc. 2003, 125, 9424−9434. (7) Mal, S. S.; Stephens, F. H.; Baker, R. T. Transition metal catalysed dehydrogenation of amine-borane fuel blends. Chem. Commun. 2011, 47, 2922−2924. (8) Neiner, D.; Karkamkar, A.; Bowden, M.; Choi, Y. J.; Luedtke, A.; Holladay, J.; Fisher, A.; Szymczak, N.; Autrey, T. Kinetic and thermodynamic investigation of hydrogen release from ethane 1,2-diamineborane. Energy Environ. Sci. 2011, 4, 4187−4193. (9) Kelly, H. C.; Edwards, J. O. Ethane 1,2-Diamineborane. J. Am. Chem. Soc. 1960, 82, 4842−4846. (10) Kelly, H. C.; Edwards, J. O. Evidence for the Open Chain Structure of Ethane 1,2-Diamineborane. Inorg. Chem. 1963, 2, 226− 227. (11) Leardini, F.; Valero-Pedraza, M. J.; Perez-Mayoral, E.; Cantelli, R.; Bañares, M. A. Thermolytic Decomposition of Ethane 1,2Diamineborane Investigated by Thermoanalytical Methods and in Situ Vibrational Spectroscopy. J. Phys. Chem. C 2014, 118, 17221−17230. (12) Carpenter, J. D.; Ault, B. S. Infrared Matrix Isolation Characterization of Aminoborane and Related Compounds. J. Phys. Chem. 1991, 95, 3502−3506. (13) Banerjee, T.; Verma, K. K.; Khanna, A. Liquid−Liquid Equilibrium for Ionic Liquid Systems Using COSMO-RS: Effect of Cation and Anion Dissociation. AIChE J. 2008, 54, 1874−1885. (14) Kumar, A. A. P.; Banerjee, T. Thiophene separation with ionic liquids for desulphurization:A quantum chemical approach. Fluid Phase Equilib. 2009, 278, 1−8. (15) Himmelberger, D. W.; Alden, L. R.; Bluhm, M. E.; Sneddon, L. R. Ammonia Borane Hydrogen Release in Ionic Liquids. Inorg. Chem. 2009, 48, 9883−9889. (16) Banerjee, B.; Kundu, D.; Pugazhenthi, G.; Banerjee, T. Quantum chemical and experimental insights for the ionic liquid facilitated thermal dehydrogenation of ethylene diamine bisborane. RSC Adv. 2015, 5, 85280−85290. (17) Bluhm, M. E.; Bradley, M. G.; Butterick, R., III; Kusari, U.; Sneddon, L. G. Amineborane-Based Chemical Hydrogen Storage: Enhanced Ammonia Borane Dehydrogenation in Ionic Liquids. J. Am. Chem. Soc. 2006, 128, 7748−7749. (18) Sahler, S.; Konnerth, H.; Knoblauch, N.; Prechtl, M. H. G. Hydrogen storage in amine boranes: Ionic liquid supported thermal dehydrogenation of ethylene diamine bisborane. Int. J. Hydrogen Energy 2013, 38, 3283−3290. (19) (a) Lin, S. T.; Sandler, S. I. A Priori Phase Equilibrium Prediction from a Segment Contribution Solvation Model. Ind. Eng. Chem. Res. 2002, 41, 899−913. (b) Burnett, R. I. Predicting liquidphase thermodynamic properties using COSMO-SAC. Ph.D. Thesis, University of Delaware, USA, 2012. (20) Klamt, A. Conductor-like Screening Model for Real Solvents: A New Approach to the Quantitative Calculation of Solvation Phenomena. J. Phys. Chem. 1995, 99, 2224−2235. (21) Zhang, H.; Wu, J.; Zhang, J.; He, J. 1-Allyl-3-Methylimidazolium Chloride Room Temperature Ionic Liquid: A New and Powerful Nonderivatizing Solvent for Cellulose. Macromolecules 2005, 38, 8272−8277. (22) Mizumo, T.; Marwanta, E.; Matsumi, N.; Ohno, H. Allylimidazolium Halides as Novel Room Temperature Ionic Liquids. Chem. Lett. 2004, 33, 1360−1361. M


Experimental Insights into the Thermal Dehydrogenation of Ethylene

Recommend Documents